Pathobiological predictors of behavior in invasive lobular carcinoma of the breast

17 3.3 IMMUNO-MARKER EXPRESSION AND ITS ASSOCIATION WITH CLINICOPATHOLOGICAL 3.4 ASSOCIATION OF HISTOLOGIC TYPE WITH CLINICOPATHOLOGICAL CHARACTERISTICS AND... Pleomorphic variant was

PATHOBIOLOGICAL PREDICTORS OF BEHAVIOUR IN INVASIVE

LOBULAR CARCINOMA OF THE BREAST

CHEOK POH YIAN

ACKNOWLEDGEMENTS

During my graduate studies, several persons and institutions collaborated directly and indirectly with my research Without their support it would be impossible for me to finish my work and I would like to dedicate this section to recognise their support

I want to start expressing my sincere acknowledgement to my supervisor, Associate Professor Tan Puay Hoon, Head of Pathology Department, Singapore General Hospital (SGH) because she gave me the opportunity to research under her guidance and supervision I received motivation, encouragement and support from her during my candidature With her, I have learned how to bring my ideas across effectively

My heart-felt appreciation to Professor Bay Boon Huat, Head of Department of Anatomy, Yong Loo Lin School of Medicine, National University of Singapore (NUS) for his encouragement and valuable suggestions

I also want to thank the motivation and support I received from Dr Aye Aye Thike I am completely grateful for her guidance and knowledge in helping me complete my work

I would like to express my sincere thanks to all staff and students of the Department of Anatomy NUS, and Department of Histopathology SGH, for creating such an excellent environment for research and friendship

The Grant from Singapore Cancer Syndicate MS04 provided the funding and the resources for the progress of this research Last, but most importantly, I would like to thank my family, for their unconditional support, inspiration and love

TABLE OF CONTENTS

ACKNOWLEDGEMENTS I SUMMARY V LIST OF TABLES VII LIST OF FIGURES IX

1 INTRODUCTION 1

1.4 BACKGROUND ON INVASIVE LOBULAR CARCINOMA (ILC) OF THE BREAST 8

2 MATERIALS AND METHODS 17

3.3 IMMUNO-MARKER EXPRESSION AND ITS ASSOCIATION WITH CLINICOPATHOLOGICAL

3.4 ASSOCIATION OF HISTOLOGIC TYPE WITH CLINICOPATHOLOGICAL CHARACTERISTICS AND

3.5 ASSOCIATION OF PLEOMORPHIC VARIANT WITH CLINICOPATHOLOGICAL CHARACTERISTICS AND

3.11 PATIENTS’ OUTCOME : UNIVARIATE AND MULTIVARIATE ANALYSES 91

4 DISCUSSION 97

4.1 SIGNIFICANCE OF CLINICOPATHOLOGICAL CHARACTERISTICS 97

SUMMARY

Invasive lobular carcinoma (ILC) accounts for approximately 10% of invasive breast carcinoma and its incidence appears to be increasing especially amongst post-menopausal women Morphologically, ILC is characterised by cells that are bland in appearance, have scant cytoplasm and infiltrate the stroma in single files Probably due

to its diffuse infiltrative growth pattern, ILC tends to be less discrete when presenting as

a breast lump Radiological studies in early diagnosis can be challenging as it tends to permeate imperceptibly through the breast stroma, thus leading to often occult mammographic appearances

ILC is the second most common histologic type of breast cancer and its incidence is reported to be lower in Asian countries compared to the western population Studies on the clinical outcome and prognostic characteristics of ILC have been few in the Asian population, therefore, warranting a detailed study of their clinical features and outcome

in the Singapore population In this study, the clinicopathological characteristics and immunohistochemical profile of ILC in a large series of Singaporean women were assessed, including its association with triple negativity and basal phenotype Using immuno-markers Estrogen Receptor (ER), Progesterone Receptor (PR), HER-2, Mammaglobin, Ki-67, Cytokeratin High Molecular Weight (CK HMW), Cytokeratin 14 (CK14) and Epidermal Growth Factor Receptor (EGFR), this study investigated the differences in characteristics and outcome between ILC and mixed ILC/invasive ductal

carcinoma (IDC), the pleomorphic and non-pleomorphic variants of ILC, triple negative ILC and non-triple negative ILC and lastly between basal-like ILC and non basal-like ILC

In these analyses, mixed ILC/IDC was associated with higher histologic grade, lobular variant, absence of associated lobular carcinoma in-situ and HER-2 positivity Pleomorphic variant was associated with higher histologic grade, increased proliferative activity, positive EGFR status, negative ER and PR status Triple negativity was associated with older age, higher histologic grade, the pleomorphic variant, negativity for CK HMW and Mammaglobin Basal phenotype was defined as expression of at least one of the two immuno-markers CK14 or EGFR This phenotype was associated with older age and presence of accompanying LCIS Molecular classification using surrogate immunohistochemical markers ER, PR and HER-2 revealed the HER-2 overexpressing molecular subtype to have the worst disease -free outcome

tubulo-Similar to other Asian countries, incidence of ILC is relatively low in Singapore The pleomorphic variant, triple negativity and the basal phenotype in ILC were associated with worse characteristics but have no impact with regard to patient survival Some of the clinicopathological parameters have been re-emphasized to predict patient outcome and in this study, tumour size, histologic grade and lymph node status remained as

LIST OF TABLES

TABLE 1.2.1 WHO CLASSIFICATION OF BREAST CANCER 3 TABLE 1.3.1 LEADING CANCER SITES OF NEW CASES AND DEATHS WORLDWIDE 5 TABLE 1.3.2 LEADING CANCER SITES OF DEATHS IN SINGAPORE FEMALES FROM 2003-

2007 6 TABLE 1.3.3 PROPORTION OF ILC DIAGNOSED IN DIFFERENT COUNTRIES 7 TABLE 2.4.1 ANTIBODY DETAILS 21 TABLE 3.2.1 CLINICOPATHOLOGICAL CHARACTERISTICS OF THE ENTIRE SERIES (N=345) 29 TABLE 3.3.1 IMMUNO-MARKERS STATUS IN ENTIRE SERIES (N = 345) 31 TABLE 3.3.2 ASSOCIATION OF HORMONAL MARKER ER STATUS WITH

CLINICOPATHOLOGICAL CHARACTERISTICS 36 TABLE 3.3.3 ASSOCIATION OF HORMONAL MARKER PR STATUS WITH

CLINICOPATHOLOGICAL CHARACTERISTICS 37 TABLE 3.3.4 ASSOCIATION OF HER-2 STATUS WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 38 TABLE 3.3.5 ASSOCIATION OF CK HMW STATUS WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 39 TABLE 3.3.6 ASSOCIATION OF CK14 STATUS WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 40 TABLE 3.3.7 ASSOCIATION OF EGFR STATUS WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 41 TABLE 3.3.8 ASSOCIATION OF KI-67 STATUS WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 42 TABLE 3.3.9 ASSOCIATION OF MAMMAGLOBIN STATUS WITH CLINICOPATHOLOGICAL CHARACTERISTICS 43 TABLE 3.4.1 ASSOCIATION OF HISTOLOGIC TYPE WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 45 TABLE 3.4.2 ASSOCIATION OF HISTOLOGIC TYPE WITH IMMUNO-MARKERS 46 TABLE 3.5.1 DISTRIBUTION OF PLEOMORPHIC VARIANT IN THE ENTIRE SERIES 47 TABLE 3.5.2 ASSOCIATION OF PLEOMORPHIC VARIANT WITH CLINICOPATHOLOGICAL CHARACTERISTICS 48 TABLE 3.5.3 ASSOCIATION OF THE PLEOMORPHIC VARIANT WITH IMMUNO-MARKERS 50 TABLE 3.6.1 DISTRIBUTION OF TRIPLE NEGATIVITY IN THE ENTIRE SERIES 51 TABLE 3.6.2 ASSOCIATION OF TRIPLE NEGATIVITY WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 52 TABLE 3.6.3 ASSOCIATION OF TRIPLE NEGATIVITY WITH IMMUNO-MARKERS 53 TABLE 3.7.1 DISTRIBUTION OF BASAL PHENOTYPE IN THE ENTIRE SERIES ACCORDING TO DIFFERENT DEFINITION 54

TABLE 3.7.2 ASSOCIATION OF BASAL PHENOTYPE WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 55 TABLE 3.7.3 ASSOCIATION OF BASAL PHENOTYPE WITH IMMUNO-MARKERS 56 TABLE 3.8.1 CRITERIA USED FOR MOLECULAR SUBTYPE AND THE DISTRIBUTION OF MOLECULAR SUBTYPE IN THE ENTIRE SERIES 58 TABLE 3.8.2 ASSOCIATION OF MOLECULAR SUBTYPE WITH CLINICOPATHOLOGICAL CHARACTERISTICS 59 TABLE 3.8.3 ASSOCIATION OF MOLECULAR SUBTYPE WITH IMMUNO-MARKERS 60 TABLE 3.9.1 E-CADHERIN AND P120 CATENIN EXPRESSION IN ILC 61 TABLE 3.9.2 P120 CATENIN CYTOPLASMIC AND CYTOPLASMIC MEMBRANE

LOCALISATION AMONG E-CADHERIN POSITIVE AND NEGATIVE TUMOURS 63 TABLE 3.9.3 ASSOCIATION OF E-CADHERIN STATUS WITH CLINICOPATHOLOGICAL

CHARACTERISTICS 65 TABLE 3.11.1 UNIVARIATE COX REGRESSION MODEL FOR DISEASE-FREE SURVIVAL (DFS) AND OVERALL SURVIVAL (OS) ON CLINICOPATHOLOGICAL CHARACTERISTICS AND IMMUNO-MARKERS 93 TABLE 3.11.2 MULTIVARIATE COX REGRESSION MODEL FOR DISEASE-FREE SURVIVAL (DFS) AND OVERALL SURVIVAL (OS) ON CLINICOPATHOLOGICAL CHARACTERISTICS AND IMMUNO-MARKERS 95 TABLE 3.12.1 LOCOREGIONAL RECURRENCE AND DISTANT SITES OF FIRST RECURRENCE (N=83) 96 TABLE 4.1.1 CLINICOPATHOLOGICAL CHARACTERISTICS OF ILC TUMOURS FROM OTHER STUDIES 99

LIST OF FIGURES

FIGURE 1.3.1 TEN MOST FREQUENT CANCER SITES IN SINGAPOREAN WOMEN 5

FIGURE 1.3.2 INCIDENCE OF BREAST CANCER IN SINGAPOREAN FEMALE FROM 1968 TO 2007 6

FIGURE 1.4.1 (A) BENIGN LOBULES (B) CLASSIC ILC CHARACTERISED BY MONOMORPHIC CELLS THAT HAVE SCANT CYTOPLASM AND INFILTRATE THE STROMA IN SINGLE FILES 8

FIGURE 1.9.1 EXPRESSION OF E-CADHERIN IN NORMAL DUCTS AND IN ILC 15

FIGURE 2.2.1 SCHEMATIC REPRESENTATION OF TMA CONSTRUCTION AND RESULTING TMA SECTION 18

FIGURE 2.4.1 POLYMERIC METHOD 19

FIGURE 3.1.1 CLASSICAL VARIANT 24

FIGURE 3.1.2 ALVEOLAR VARIANT 25

FIGURE 3.1.3 SOLID VARIANT 26

FIGURE 3.1.4 TUBULO-LOBULAR VARIANT 26

FIGURE 3.1.5 PLEOMORPHIC VARIANT 27

FIGURE 3.3.1 IMMUNOHISTOCHEMICAL EXPRESSION OF ER, PR AND HER-2 32

FIGURE 3.3.2 IMMUNOHISTOCHEMICAL EXPRESSION OF BASAL MARKERS CK HMW, CK14 AND EGFR 33

FIGURE 3.3.3 IMMUNOHISTOCHEMICAL EXPRESSION OF KI-67 AND MAMMAGLOBIN 34

FIGURE 3.9.1 VARYING IMMUNOHISTOCHEMICAL EXPRESSION OF E-CADHERIN IN ILC 62 FIGURE 3.9.2 DIFFERENTIAL EXPRESSION OF E-CADHERIN AND P120 CATENIN IN ILC 64

FIGURE 3.10.1 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO TUMOUR SIZE 66

FIGURE 3.10.2 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO HISTOLOGIC GRADE 67

FIGURE 3.10.3 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO LVI 68

FIGURE 3.10.4 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO LYMPH NODE STATUS 68

FIGURE 3.10.5 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO MOLECULAR SUBTYPE 69

FIGURE 3.10.6 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO AGE 70

FIGURE 3.10.7 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO LCIS 70

FIGURE 3.10.8 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO HISTOLOGIC TYPE 71

FIGURE 3.10.9 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO LOBULAR VARIANT 71

FIGURE 3.10.10 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO TNBC CATEGORY 72

FIGURE 3.10.11 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO BASAL PHENOTYPE 72 FIGURE 3.10.12 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO PR STATUS 73 FIGURE 3.10.13 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO HER-2 STATUS 74 FIGURE 3.10.14 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO ER STATUS 75 FIGURE 3.10.15 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO KI-67 STATUS 75 FIGURE 3.10.16 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO

MAMMAGLOBIN STATUS 76 FIGURE 3.10.17 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO CK HMW STATUS 76 FIGURE 3.10.18 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO CK14 STATUS 77 FIGURE 3.10.19 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO EGFR STATUS 77 FIGURE 3.10.20 RELATIVE CUMULATIVE DFS OF ILC PATIENTS WITH RESPECT TO E-

CADHERIN EXPRESSION 78 FIGURE 3.10.21 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO AGE 79 FIGURE 3.10.22 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO TUMOUR SIZE 80 FIGURE 3.10.23 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO LVI 81 FIGURE 3.10.24 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO LCIS 81 FIGURE 3.10.25 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO LN

STATUS 82 FIGURE 3.10.26 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO BASAL PHENOTYPE 83 FIGURE 3.10.27 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO

MOLECULAR SUBTYPE 84 FIGURE 3.10.28 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO GRADE 85 FIGURE 3.10.29 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO LOBULAR VARIANT 85 FIGURE 3.10.30 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO PR

STATUS 86 FIGURE 3.10.31 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO ER

FIGURE 3.10.34 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO

MAMMAGLOBIN STATUS 88 FIGURE 3.10.35 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO CK HMW STATUS 89 FIGURE 3.10.36 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO CK14 STATUS 89 FIGURE 3.10.37 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO EGFR STATUS 90 FIGURE 3.10.38 RELATIVE CUMULATIVE OS OF ILC PATIENTS WITH RESPECT TO E-

CADHERIN EXPRESSION 90

in the nipple At the end of the terminal ducts are lobules which produce milk The glandular and ductal components of the breast are embedded in tissue which consist of adipose tissue (fats) and collagenous stroma This fibro-fatty matrix holds and shapes the breast

At puberty, estrogen stimulates the growth of ducts and thickening of epithelium and periductal stroma Growth hormone and glucocorticoids contribute to ductal growth Lobuloalveolar differentiation and growth during this period are enhanced primarily by insulin, progesterone, and growth hormone (Topper and Freeman 1980)

Lymph nodes play a vital role in the spread of breast cancer The axillary lymph nodes

1.2 Classification of breast cancer

The latest World Health Organization (WHO) classification of breast cancer recognises the existence of 18 histologic types of breast cancer and their variants according to Table 1.2.1 (Tavassoli and Devilee 2003) This classification includes Invasive Ductal Carcinoma- No Special Type (IDC-NST) and 17 special types IDC-NST accounts for the majority of all breast carcinomas and it makes up approximately 50-80% of all diagnosed breast cancers ILC accounts for 5-20% of all invasive breast cancers and it is the most common special type of breast cancer (Weigelt and Reis-Filho 2009)

Introduction

Table 1.2.1 WHO classification of breast cancer

WHO classification of breast cancer (2003)

Epithelial tumours

1 Invasive ductal carcinomas of no special type

• Mixed type carcinoma

• Pleomorphic carcinoma

• Carcinoma with osteoclastic giant cells

• Carcinoma with choriocarcinomatous features

• Carcinoma with melanotic features

2 Invasive lobular carcinomas

• Classical lobular carcinoma

• Alveolar lobular carcinoma

• Solid lobular carcinoma

• Pleomorphic lobular carcinoma

5 Invasive papillary carcinoma

6 Invasive cribriform carcinoma

7 Metaplastic carcinomas

• Pure epithelial metaplastic carcinomas

→Squamous cell carcinomas

→Adenocarcinoma with spindle cell metaplasia

• Solid neuroendocrine carcinoma

• Atypical carcinoid tumour

• Small cell / oat cell carcinoma

• Large cell neuroendocrine carcinoma

13 Glycogen-rich clear cell carcinoma

Introduction

1.3 Breast cancer statistics: Global and local

1.3.1 Breast cancer statistics (all types breast cancer)

Breast cancer is the most frequently diagnosed cancer in women and is the leading cause of cancer death amongst women worldwide (Table 1.3.1)(Garcia et al 2007) Although breast cancer incidence is on the rise worldwide, mortality rate from this disease has been stable or decreasing in some countries as a result of early detection and improved treatment (Garcia et al 2007)

In Singapore, breast cancer is the most frequently diagnosed cancer in females (Figure 1.3.1) Over the last 4 decades, since the Singapore Cancer Registry started collecting and reporting statistics on cancer, breast cancer incidence has been increasing (Figure 1.3.2) It is also the top cause of cancer mortality in Singaporean females (Table 1.3.2)

Introduction

Table 1.3.1 Leading cancer sites of new cases and deaths worldwide

Estimated numbers were taken from Global cancer facts & figures 2007 (Garcia et al

Introduction

Figure 1.3.2 Incidence of breast cancer in Singaporean female from 1968 to 2007

Taken from Trends in cancer incidence in Singapore 2003-2007 with permission from

National Registry of Diseases Office

Table 1.3.2 Leading cancer sites of deaths in Singapore females from 2003-2007

Taken from Trends in cancer incidence in Singapore 2003-2007 with permission from

National Registry of Diseases Office

Introduction

1.3.2 Breast cancer statistics on ILC

Incidence of ILC ranges as low as 1-4% to as high as 5-10% in different regions and dependent on how restrictive the diagnostic criteria are Among the Asian countries, Korea reported its incidence as 2.8% from 2001 to 2008 (Jung et al 2010) and a Japanese clinical study consisting of 549 cases over 16 years found their incidence to be 1.3% (Tanaka et al 1987) The breast centre at the Baylor College of Medicine reported 8.2% of breast cancer diagnosed as ILC between 1970 and 1998 (Arpino et al 2004) A meta-analysis of 15 Internal Breast Cancer Group trials between 1978 and 2002, totalling 1,206 subjects, had 6.2% of breast cancer classified as ILC (Pestalozzi et al 2008)

In Singapore, the incidence of ILC was on the low end of the range From 2003-2007, ILC comprised 3.9% of all breast cancers (Singapore Cancer Registry 2009) From 1994-2008, the proportion of pure ILC diagnosed is 6.3% and 2.3% for mixed ILC/IDC in the department of Pathology, Singapore General Hospital (SGH)

Table 1.3.3 Proportion of ILC diagnosed in different countries

Country Proportion of ILC Year Source

USA (Baylor College of Medicine) 8.2% 1970-1998 Arpino et al., 2004

Introduction

1.4 Background on Invasive Lobular Carcinoma (ILC) of the breast

Invasive lobular carcinoma (ILC) of the breast was first fully described and established in

1941 (Foote and Stewart 1941) The in-situ components were described as having uniform cells with non-hyperchromatic nuclei and disorderly arrangement, loosely displaced towards the lumen of the terminal ducts Their invasive counterparts were described as uniformly sized cells, disorientated arrangement in a circumferential manner around ducts and lobules (targetoid growth) (Figure 1.4.1) Foote and Stewart believed that these cells were indicative of malignancy and had to be radically treated (Foote and Stewart 1941) Their criteria for reporting of ILC have been widely accepted,

in particular, the classical variant of ILC

Figure 1.4.1 (A) Benign lobules (B) Classic ILC characterised by monomorphic cells that have scant cytoplasm and infiltrate the stroma in single files

Introduction

ILC has clinicopathological characteristics that are different from invasive ductal carcinoma (IDC) Large population-based studies seen the incidence of IDC kept relatively constant while the incidence of ILC and mixed ILC/IDC have increased over the years (Verkooijen et al 2003;Li et al 2003) ILC is often more difficult to detect at an earlier stage due to its reduced tendency to form palpable masses This may be attributed to its linear pattern of infiltration eliciting little desmoplastic stromal response There are contradicting findings on prognosis of ILC, with many studies reporting better outcome compared to IDC (Korhonen et al 2004;Arpino et al 2004), while a study with long term follow-up reported the contrary (Pestalozzi et al 2008)

1.5 Histologic variants of ILC

The definition of ILC was broadened to include other growth patterns of ILC In these variants, tumour cells have similar cytologic characteristics as the classical variant but lacks the linear growth pattern (Fechner 1975;Shousha et al 1986;Fisher et al 1977;Eusebi et al 1992;Buchanan et al 2008) Several variants had been described including alveolar, solid, tubulo-lobular and pleomorphic variants The pleomorphic variant, in particular, has been shown to be associated with more aggressive clinical behaviour and worse clinical outcome A study comparing pleomorphic lobular

Introduction

consistently shown to have less favourable outcome when compared to patients harbouring tumours of the classic ILC variant (Weidner and Semple 1992;Arpino et al 2004)

1.6 Clinical features of ILC

ILC are often diagnosed in older patients compared to IDC (Pestalozzi et al Garau et al 1996;Albrektsen et al 2010) In a Norwegian study of association of histologic types with reproductive trend, proportion of IDC remained constant across age groups while proportion of ILC increased markedly with increasing age (Karl N Krecke 1983)

2008;Sastre-Radiological studies in early diagnosis of ILC can be challenging as it tends to permeate imperceptibly through the breast stroma It is less likely to be associated with calcification and its low opacity may contribute to more difficulty in detecting ILC (Yeatman et al 1995), hence a basis for its underestimation on mammography compared to IDC (Chen et al 2002) ILC have been reported to be bigger tumours compared to IDC, with increasing number of metastatic lymph nodes associated with tumour size (Yeatman et al 1995) It has also been shown to be associated with multicentricity and bilaterality (Yeatman et al 1995)

Introduction

ILC have been reported to have higher incidence of positive margin after breast conservation surgery (Santiago et al 2005) Breast conserving surgery on ILC can be challenging as it is often associated with more false-negative margins resulting in a higher frequency of conversion to mastectomy subsequently (Moore et al 2000) Yet, comparing breast conservation surgery for early stage ILC and IDC, there seems to be no significant difference in 10 year overall survival, recurrence and disease-free status There is also no difference in risk of developing contralateral breast carcinoma (Peiro et

al 2000;Kelsey et al 1993)

1.7 Risk Factors of ILC

Being female without a doubt puts one at risk of developing breast cancer established familial mutations such as BRCA1 and BRCA2 mutations are widely-known inheritable susceptibility genes for breast cancer High breast tissue density (a mammographic measure of the amount of glandular tissue relative to fatty tissue in the breast) is associated with higher breast cancer risk Biopsy confirmed hyperplasia of breast tissue especially atypical hyperplasia, and high-dose radiation to the chest as a result of medical procedures are also risk factors Epidemiologic studies have shown that reproductive and lifestyle exposures are predictive of subsequent breast cancer risk

Introduction

Other behavioural and lifestyle risk factors, particularly relevant after menopause, include physical inactivity, menopausal hormone therapy use, alcohol consumption, and high body mass index (Biglia et al 2007)

Whereas the incidence of IDC has remained stable, the incidence of ILC appears to be increasing especially amongst post-menopausal women Recent studies have suggested the association of use of hormone replacement therapy with this trend (Tanaka et al 1987;Newcomb et al 2010) Older age at first full term pregnancy and older age at menarche are significantly associated with elevated risk of ILC This risk is statistically different for ILC and IDC (Albrektsen et al 2010;Li et al 2006) Alcohol use has a more pronounced increased risk of developing lobular carcinoma, especially among postmenopausal women This risk again differs between lobular and ductal tumours (Stange et al 2006)

1.8 Molecular pathology of ILC

Breast cancer is a heterogeneous disease that arises from accumulation of complex array of genomic alterations Many studies have attempted to characterise these genomic alterations and make sensible relationship to clinical behaviour and morphology Recent genetic profiling studies have revealed unique changes in ILC at the molecular level Comparative genomic hybridisation (CGH) is a technique that allows mapping of DNA copy number changes in human tumours This technique has been

Introduction

frequently employed to characterise breast cancer Commonly reported DNA copy number changes unique to ILC include gain of 1q and 5p, and loss of 16q, 16p, 17p, 18q12–q21, and 22q (Loveday et al 2000;Günther et al 2001;Zhao et al 2004)

Genome-wide expression profiling techniques have identified abnormal gene regulations in ILC Downregulation of E-cadherin reflects results of CGH analysis where loss of genetic material in the region of 16q chromosome corresponds to CDH1 gene resulting in its low transcription and expression Genes related to basal epithelial cell markers (e.g., KRT 5, KRT 17, and EGFR) are also identified to be downregulated in ILC (Weigelt et al 2010) Other differential expressions include downregulation of genes involved in DNA repair, proliferation/cell cycle activities and up-regulation of genes

involved in lipid/prostaglandin biosysnthesis and cell migration (Simpson et al 2008)

Pleomorphic ILC was shown to be more similar to ILC than IDC at the genomic level Further accumulation of genomic abnormality is associated with the pleomorphic variant and this may explain the aggressive nature of ILC Loss of BRCA2 is reported at a higher proportion in pleomorphic ILC (Berx et al 1996)

1.9 E-cadherin and p120 catenin

Introduction

lobular cancer cells and cancer progression by increasing proliferation, invasion and metastasis Negative E-cadherin expression is one of the major defining features of lobular tumours; rather than a prognostic factor to differentiate between the outcomes

of ductal and lobular tumours (Coradini et al 2002) Lobular carcinomas have diminished, absent or aberrant E-cadherin expression, while presence of complete membrane staining indicates ductal phenotype

p120 catenin belongs to a group of proteins called catenins and it is attached to the juxtamembrane domain of E-cadherin in the intracellular cytoplasm p120 catenin’s expression in the cell membrane is an indication that E-cadherin is intact in the cell membrane, while its localisation to the cytoplasm suggests E-cadherin is non-functional

or absent Cytoplasmic localisation of p120 catenin is a positive marker for lobular neoplasia, from the early stage of atypical lobular hyperplasia to invasive lobular carcinoma (Sarrió et al 2004) It has been reported to be helpful in diagnosis of metastatic lobular carcinoma (Dabbs et al 2007) A positive p120 marker may be easier

to interpret in some instances compared to negative E-cadherin expression, especially when E-cadherin is also expressed in myoepithelial cell membranes, hence presenting a challenge to interpret precisely which cells are positive with E-cadherin

Introduction

Figure 1.9.1 Expression of E-cadherin in normal ducts and in ILC

(A) Presence of complete E-cadherin membrane staining in normal ductal component (B) E-cadherin negative expression in ILC

1.10 Scope of study

Studies on the clinical outcome and prognostically significant characteristics of ILC have been few in the Asian population This warrants a detailed study of their clinicopathological features and outcome in the Singapore population including their association with triple negativity and basal phenotype I propose to document more comprehensively the clinicopathological characteristics and immunohistochemical profile of ILC of in a large series of affected Singaporean women

Hypothesis:

Introduction

that exhibit basal-like characteristics and the pleomorphic morphology of tumour cells are predictors that are associated with poor patient outcome This current work hypothesizes that a subset of lobular cancers which exhibit these characteristics will have worse outcome

The objectives of this study are as follows:

1 To establish prognostic and predictive values of clinicopathological characteristics and immuo-markers in ILC The immuno-markers to be assessed are ER, PR,

HER-2, Mammaglobin, Ki-67, CK HMW, CK14 and EGFR

2 To investigate the differences in characteristics and outcome of ILC and mixed ILC/IDC

3 To determine the significance of the pleomorphic variant of ILC

4 To evaluate the existence and significance of triple negativity in ILC

5 To ascertain the existence and significance of basal phenotype in ILC

6 To determine the significance of molecular subtyping using surrogate markers ER, PR and HER-2

7 To survey the usefulness of E-cadherin and p120 expression in interpretation of ILC

by documenting the frequency of aberrant E-cadherin staining pattern, p120 catenin cytoplasmic expression and determine the association of E-cadherin expression pattern with outcome

Materials and methods

2 Materials and Methods

2.1 Study Population

The study population was derived from the database in the archives of the Pathology Department of Singapore General Hospital (SGH) All patients with invasive lobular carcinoma of the breast diagnosed at the Pathology Department (SGH) were identified

A total of 669 cases, including pure ILC and mixed ILC/IDC histologic types were reported Patients who did not have archival materials available in the department were excluded from the study Haematoxylin and Eosin (H&E) stain and E-cadherin immunohistochemistry were preformed on whole tumour sections Tumours that did not express E-cadherin or tumours where E-cadherin expression was aberrant were included in the study and representative tumour areas were selected for Tissue Micro-Array (TMA) construction A final of 345 cases from 1994 to 2008 were used in this study Centralised Institutional Review Board approval was obtained

2.2 Tissue Micro-Array (TMA) Construction

In this study, representative areas of tumour were circled on H&E sections Only tumour

Materials and methods

expression for E-cadherin 1mm core and 2 cores per case were used for TMA construction as illustrated in Figure 2.2.1

Figure 2.2.1 Schematic representation of TMA construction and resulting TMA section

(A) Beecher Tissue microarrayer (B) Tissue cores are removed from the ‘donor’ block and inserted into premade holes of the ‘recipient’ block Microtomes are used to cut TMA sections (C) Overview of a H&E stained TMA section (D) Magnification of H&E spot and immunohistochemistry spot

Materials and methods

2.3 Patient’s Clinicopathological Characteristics

Patients’ clinicopathological history and tumour characteristics including age, ethnicity, tumour size, histological grade, lobular variant, accompanying lobular carcinoma in-situ (LCIS), lymphovascular invasion (LVI) and lymph node (LN) status were obtained from the database H&E tumour sections were reviewed and the tumours were classified as classic, alveolar, solid, tubulo-lobular, pleomorphic variant or mixed type

2.4 Immunohistochemistry

Immunohistochemistry was performed by the polymeric-based two-step method This method consists of a compact polymer to which multiple molecules of enzyme (to catalyse substrate for visualisation) and the secondary antibody (specific for the primary antibody) are attached (Figure 2.4.1)

Primary antibody Secondary antibody Antigen

Dextran polymer

Materials and methods

Immunohistochemical assays were performed on formalin-fixed paraffin embedded sections Four µm thick sections were cut from the TMA blocks, mounted on silanized glass slides and dried on heating bench for at least 20 minutes The sections were deparaffinised in 2 changes of xylene for 2 min each This was followed by rehydration

in decreasing concentration of alcohol from 100% followed by 95% to 75% and finally in water The rehydrated slides were subjected to antigen retrieval according to Table 2.4.1 For heat-induced antigen retrieval, slides were heated in 0.01M Tris-0.001M EDTA pH9 antigen retrieval solution for 15 minutes at sub-boiling temperature of 98°C in a microwave (Milestone T/T mega) For enzymatic antigen retrieval, sections were incubated with protease for 10 minutes at 40ºC Slides were then run on Dako Autostainer Plus according to the following steps: Endogenous peroxidase activity was blocked using hydrogen peroxide (Dako S2022) for 10 minutes Slides were incubated with the respective optimally diluted primary antibody for 30 minutes at room temperature Detection was achieved using Dako Envision Detection kit (Dako K5007) which is a dextran backbone conjugated with secondary antibodies to mouse or rabbit immunoglobulin and horseradish peroxidase (Figure 2.4.1) Addition of substrate chromogen, diaminobenzidine (DAB) for 5 minutes will produce an insoluble brown precipitate catalysed by HRP

Slides were removed from the autostainer and countered stained with Mayer’s Haematoxylin (Dako S3309) for 1 minute and coverslipped Appropriate controls were run with each batch of slides

Materials and methods

Table 2.4.1 Antibody Details

Primary

Antibody Dilution

Antigen Retrieval

Monoclonal

Thermo Scientific #RM-9106-S

Materials and methods

cut-offs were used for established prognostic immuno-markers For ER and PR status, at least 1% of tumour cells have to display a minimum of 1+ nuclear staining to be considered positive, according to the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines (Hammond et al 2010) For HER-2 positive status, more than 30% of tumour cells have to exhibit 3+ uniform intense cytoplasmic membrane staining, according to the ASCO/CAP guidelines ((Wolff et al 2007) Ki-67 high expression was defined as ≥10% of positive tumour cells E-cadherin membranous expression of at least 10% of tumour cells was considered positive for prognostic comparison For the remaining immuno-markers (CK HMW, CK14, EGFR, mammglobin and p120 catenin), cut-off of at least 1% positive tumour cells stained defined positive status

2.5 Statistical analysis

Associations between categorical variables and status of immuno-markers were assessed using Chi-square and Fisher’s exact tests Survival data including survival time, disease-free interval and development of distant metastasis were retrieved from the database The primary endpoints of this study were Disease-Free Survival (DFS) and Overall Survival (OS) DFS was defined as the length of time between date of diagnosis and first observation of disease recurrence, either loco-regional recurrence (including ipsilateral and contralateral breast recurrence) or distant metastasis censored at time of last follow-up or death from any cause OS was defined as the interval between date of

Materials and methods

diagnosis and death from all causes Survival plots according to selected tumour characteristics and immuno-marker status were drawn using the Kaplan-Meier method The log rank test was used to assess survival differences between strata Univariate and multivariate Cox proportional hazards regression analyses were used to assess the independent prognostic significance of clinicopathological parameters and immuno-markers on survival Initially, a univariate analysis was performed and parameters identified as significant were included in a multivariate analysis Parameters included in the analyses were age, tumour size, histologic grade, accompanying LCIS, LVI, LN status Hazard ratio was presented with 95% confidence intervals Analyses were performed

using Statistical Package for the Social Science (SPSS), version 18 All p-values are sided with p-value significance at <0.05

“targetoid” appearance

Figure 3.1.1 Classical variant

Figure 3.1.4 Tubulo-lobular variant

Figure 3.1.5 Pleomorphic variant

Results

3.2 Patients and tumour characteristics

A final number of 345 patients with pure ILC and mixed ILC/IDC were included for analysis Table 3.2.1 summarised the clinical and histopathological features for the entire series Out of these cases, 246 were pure ILCs and 99 were mixed ILC/IDCs The age ranged from 25-86 years, mean age was 55 years Majority of patients were Chinese (80.6%), Malays made up 6.7%, Indians 5.5% and other ethnicities 5.2% of the series There were 5 (1.4%) cases without ethnicity records Tumour size ranged from 1mm to 130mm, mean measuring 31 mm There were more patients with T2 stage tumours (tumour size larger than 20mm) at 55.1%, compared with 38.0% with T1 stage This series comprised 97 (28.1%) grade 1 tumours, 195 (56.5%) grade 2 tumours and 39 (11.3%) grade 3 tumours The most frequently detected lobular variant was the classical variant (31.3%), followed by solid (13%), pleomorphic (12.5%), alveolar (9.3%) and tubulo-lobular variant (8.4%) The remaining 25.5% tumours were of mixed subtypes, they mostly comprised classical mixed with other variants Lymphovascular invasion (LVI) was detected in 68 patients (19.7%) and accompanying LCIS was present in 176 tumours (51.0%) Axillary lymph node metastases were detected in 139 (40.3%) out of 345 patients, 52 (15.1%) of them had 1 to 3 metastatic lymph nodes (pN1), 38 (11.0%) had 4

to 9 metastatic lymph nodes (pN2) and 49 (14.2%) had more than 9 metastatic lymph nodes (pN3)